Pricing

Buy credits. Choose where to spend them.

You are not buying an API. You are buying access to knowledge infrastructure — one wallet of Orves Credits, consumed by whichever products you use, in whatever mix your architecture needs.

How it works

One wallet. Two meters.

01

Load the wallet

One balance of Orves Credits — no per-product currencies, no fifteen-dimension bills. Start free with 15,000 credits.

02

Consume anywhere

Each product meters in its own natural unit — pages and knowledge operations. The wallet is shared; the consumption is not.

03

Audit every debit

An immutable ledger reconciles every credit to the unit of work that consumed it.

Consumption

What a credit buys, per product.

ProductUnit of consumptionCreditsStatus
Parser 1 page 1 credit draft ratio
Verifiable Knowledge 1 page canonicalized 2 credits draft ratio
1 knowledge operation — ingest · update · retrieve · consolidate 5 credits draft ratio
Draft ratios — honest by design. Final rates are published at launch, derived from measured unit costs. Certificates are never billed: every artifact ships certified, and public verification is free forever.

Ratios follow the value chain: the closer a unit sits to provable knowledge, the more it delivers — canonicalizing a page is worth more than parsing it, and a knowledge operation more still. Verifiable Knowledge never charges per token — it meters knowledge operations, because it sells infrastructure, not inference. Continuous monitoring is recurring — the Living Knowledge solution is sized by watched sources, monitored documents and update frequency.

Bundles

Not discounts — adoption stages.

Start where your problem starts: structured evidence, provable knowledge, or knowledge that never goes stale.

Structured Evidence

You already have AI.

Parser

High-fidelity structured evidence into your own pipelines and AI systems — per page, no lock-in.

Provable Knowledge

You need knowledge you can trust.

Parser + Verifiable Knowledge

Raw sources become a governed knowledge base — origin, versions, evidence — served to any model, agent or person.

Living Knowledge

Your knowledge changes daily.

Parser + Verifiable Knowledge + continuous monitoring

The full loop: sources watched, change governed, knowledge that never quietly goes stale.

Land on one product. Expand when the next problem appears. Bundle terms published with rates.

Teams & Enterprise

Collaboration and governance, separated from consumption.

Workspaces for how you work.

Members, spaces, permissions, audit, SSO/SCIM, retention, freshness policies and SLAs — workspace terms cover how you work; credits cover the work performed.

members & permissions in development SSO · SCIM · RBAC planned freshness policies & SLAs Talk to us
Mechanics

Billing you can audit — by design.

Immutable ledger

Every debit an append-only entry. Reconcile to the unit.

Reserve, then settle

Work reserves credits and settles on completion. No surprises.

Idempotent charging

Retries never double-bill — enforced server-side.

Certificates never billed

Verification is part of the infrastructure, not a meter. Public endpoint: free forever, by design.

No dollar rates yet — on purpose. Draft credit ratios are published above; currency rates appear at launch, derived from measured unit costs. Pilot partners get founding terms.
Questions

FAQ

Why one wallet instead of per-product pricing?

Because you are buying access to one infrastructure, not a bundle of APIs. A single balance of Orves Credits removes the friction of deciding budgets per product before you know your mix.

Why does Verifiable Knowledge not charge per token?

Charging tokens would put it in the same category as the models. It sells persistent infrastructure, so it meters knowledge operations — ingest, update, retrieve, consolidate — not words generated.

Do external model costs come through Orves?

The design separates Orves usage from third-party LLM costs — pass-through or bring-your-own-key where it makes sense. Final mechanics published with rates.

Why are there no dollar prices on this page?

Because we have not finished measuring unit costs, and we do not publish numbers we cannot defend. The same discipline as our benchmarks.

Can I start small?

Yes — 15,000 free credits, one product, one problem. Grow when it earns it.

More questions → FAQ

Build AI systems that know where every answer came from.

Free sandbox — real API, sample corpus, verify a real certificate. Buy credits when it earns it.